Agitation and misunderstanding of the similar concepts, which happens quite often with new and emerging philosophies in Information Systems field, can be quite problematic for scientists and industry professionals and drag them into undesired direction out of their focus; consequently consuming additional resources, producing unwanted or biased products, and learning on conclusions based on false assumptions.
To avoid such issues and problems, I consulted more than 50 academic sources (scientific journal articles and conference papers) and produced a simple Framework (figure below) to differentiate elements from the cluster of vague concepts concerned with analysing massive data. The concepts focused in this framework are: Business Intelligence, Big Data, Data Analytics and Knowledge Discovery.
As presented in Figure above, we see Knowledge Discovery as the highest concept, which in addition to other methods includes Data Analytics to discover or produce new knowledge. Further on, we see Data Analytics as larger entity encompassing various disciplines, including Big Data Analytics and Business Intelligence.
The concept of the Big Data is generally seen as a part of Big Data Analytics. Taking into account intention, purpose and underlying business philosophies, we see Big Data Analytics and Business Intelligence at the same level. However, taking into account technical structure, relevant software applications and data, we see Big Data and Business Intelligence as the concepts at the same level as well.
We also see a data focus as the major difference between BI and Big Data. Big Data encompass unstructured, semi-structured and structured data, however main focus is on unstructured data, while focus of BI is on structured data. While BI requires DW and/or data marts to support reporting, Big Data can work with DW, but is not required. In reports based on BI system, there is a requirement to have structured master and transactional data. For example, to use or analyse sales transactional data in meaningful and understandable way, we must have master data describing the properties of sales (such as store, location or product descriptions). Big Data concept is not subject of those requirements. For example, we dont need structured data to analyse content of respective emails or to analyse appleals of invidivuals submitted to public administration institutions.
Update: 28th July 2017 – This post is based on my research paper published by Lecture Notes in Business Information Processing, vol 285. Springer.
DOI: https://doi.org/10.1007/978-3-319-58801-8_10